how extract table RowNames based on table elements

Hello all,

I have a table of only one colume with RowNames.

The table have numerical values and NaN.

I want to extract the RowNames of all of the NaN in the table. How can I do that?

Thanks in advance!

Answer 1

[Voss]

% a table as you describe
t = table([1;NaN;3;4;NaN],'RowNames',{'A';'B';'C';'D';'E'})

t = 5×1 table
         Var1
         ____

    A      1 
    B    NaN 
    C      3 
    D      4 
    E    NaN 

% extract RowNames of the NaN rows
nan_rows = t.Properties.RowNames(isnan(t{:,:}))

nan_rows = 2×1 cell array
    {'B'}
    {'E'}

Answer 2

[Steven Lord]

I'd use ismissing, to avoid the need to extract the whole contents of the table into a separate array.

% a table as you describe
t = table([1;NaN;3;4;NaN],'RowNames',{'A';'B';'C';'D';'E'})

t = 5×1 table
         Var1
         ____

    A      1 
    B    NaN 
    C      3 
    D      4 
    E    NaN 

% extract RowNames of the NaN rows
nan_rows = t.Properties.RowNames(ismissing(t))

nan_rows = 2×1 cell array
    {'B'}
    {'E'}

This also works if t has variables of different types. Yes, I could have listed "elderberry" but I need a missing value for the example.

t.FruitNames = ["apple"; "banana"; "cherry"; "date"; missing]

t = 5×2 table
         Var1    FruitNames
         ____    __________

    A      1     "apple"   
    B    NaN     "banana"  
    C      3     "cherry"  
    D      4     "date"    
    E    NaN     <missing> 

ismissing(t)

ans = 5×2 logical array
   0   0
   1   0
   0   0
   0   0
   1   1

rowWithMissingNumberAndFruit = t.Properties.RowNames(all(ismissing(t), 2))

rowWithMissingNumberAndFruit = 1×1 cell array
    {'E'}

rowWithMissingNumberOrFruit = t.Properties.RowNames(any(ismissing(t), 2))

rowWithMissingNumberOrFruit = 2×1 cell array
    {'B'}
    {'E'}

The approach Voss posted wouldn't work here.

isnan(t{:,:})

Incorrect number or types of inputs or outputs for function 'isnan'.